---

Session 2: Visualizing Tabular data

Multivariate Case

Course: Visual Analytics for Policy and Management - Exercise 3 Multivar

Hanson Shi

PART1 - Descriptive plots

library(openxlsx)
library(reshape)
library(ggplot2)
link2="https://github.com/EvansDataScience/data/raw/master/safeCitiesIndex.xlsx"
safe2=read.xlsx(link2)
safe2A=melt(safe2,id.vars = 'city')

base  = ggplot(safe2A, aes(x = variable, y = value, group = city))

plot1 = base + geom_polygon(fill = 'darkgrey',
                            col='darkorange') + 
               coord_polar()

plot2 = plot1 + facet_wrap(~reorder(city,value, median, order=TRUE),
                           ncol = 7)

plot3 = plot2 + theme(axis.text.x = element_text(size = 6, colour = 'black'),
                legend.position = 'top',
                strip.text = element_text(size = 20))
brdBkgnGrid = element_rect(fill = "white",colour = "grey",
                           size = 2,
                           linetype = "dashed")

lineGrid = element_line(size = 0.4,
                        linetype = 'solid',
                        colour = "steel blue")

plot3 + theme(panel.background = brdBkgnGrid,
              panel.grid.major = lineGrid,
              plot.caption = element_text(hjust = 0),
              plot.title = element_text(hjust = 0.5)) +
        labs(title='The safety levels of cities in the world', 
             x="variables", 
             y="values",
             caption="Source: Economist - worldbook") 

PART2 - Inferential plots

library(dotwhisker)
library(broom)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following object is masked from 'package:reshape':
## 
##     rename

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(margins)

cut=median(safe2$PERSONAL)
safe2$PERSONAL_bi=ifelse(safe2$PERSONAL>cut,
                         1,0)
logit_PERSONAL = glm(PERSONAL_bi~ .,
                          data = safe2[,-c(1,5)],
                          family = "binomial")

margins_logit_PERSONAL = margins(logit_PERSONAL) 

marginalSummary=summary(margins_logit_PERSONAL)

base = ggplot(data = marginalSummary)

eff1= base +  geom_point(aes(factor, AME), 
                         alpha=0.5)

eff2= eff1 + geom_errorbar(aes(x = factor, 
                               ymin = lower, 
                               ymax = upper),
                           colour = c('steelblue','dark red','dark red'))

MESSAGE1="increasing on average \n 1.7% the probability \n of being a safe city"

eff3 = eff2 + geom_hline(yintercept = 0, 
                         linetype= 'dashed', 
                         size= 0.5, 
                         alpha= 0.7) + 
              theme_minimal()
             
eff3 + annotate("text", 
                x = 1.4, 
                y = 0.02, 
                label = MESSAGE1) +
       labs(title='Marginal effect of coefficients values', 
            x="Category", 
            y="Average Marginal Effect",
            caption='Source: Economist - worldbook') +
       theme(plot.caption = element_text(hjust = 0),
             plot.title = element_text(hjust = 0.5)) +
       annotate("text", 
                x = 0.7, 
                y = -0.002, 
                label = 'reference line') +
       annotate("text", 
                x = 2.22, 
                y = -0.012, 
                label = 'lower value') +
       annotate("text", 
                x = 3.22, 
                y = 0.01, 
                label = 'upper value')